Clinical validation of the short and long UNESP-Botucatu scales for feline pain assessment

Introduction

Being free from pain, injury, and disease is one of the fundamental five freedoms in animal welfare (Mellor, 2016; Robertson, 2018). The International Association for the Study of Pain (IASP) defines pain as “an unpleasant sensory and emotional experience associated with, or resembling that associated with, actual or potential tissue damage”. In addition, the “inability to communicate does not negate the possibility that a human or a nonhuman animal experiences pain” (Raja et al., 2020).

Pain management in animals has improved over time due to significant advances in its recognition, assessment and treatment (Lorena et al., 2013; Lorena et al., 2014; Robertson, 2018). However, animals still do not receive adequate analgesic therapy due to the lack of validated pain assessment scoring systems. Indeed, many veterinarians feel insecure and unable to correctly identify the presence of pain in animals (Lorena et al., 2013; Lorena et al., 2014).

In cats, the development and validation of species-specific, composite, and multidimensional pain scales have influenced feline pain recognition and assessment. These instruments may also provide clinical decision aid for analgesic provision once reaches a certain threshold (i.e., rescue analgesia) (Brondani et al., 2012; Brondani et al., 2013a; Brondani et al., 2013b; Merola & Mills, 2016; Reid et al., 2017; Steagall & Monteiro, 2019; Evangelista et al., 2019). They often include physiological and behavioral parameters associated with the observational and dynamic interaction between the cat and the observer.

There are currently three scales with reported validation for evaluating pain in cats: the UNESP-Botucatu multidimensional pain assessment scale (UFEPS) (Brondani, Luna & Padovani, 2011; Brondani et al., 2012; Brondani et al., 2013a; Brondani et al., 2013b), the Glasgow Feline Composite Measure Pain Scale (Glasgow CMPS-Feline) (Calvo et al., 2014; Reid et al., 2017), and the Feline Grimace Scale (Evangelista et al., 2019; Evangelista et al., 2020; Watanabe et al., 2020).

The UFEPS is the only one with reported validation (i.e., construct, content and criterion validation, reliability and sensitivity) in several languages such as Portuguese (Brondani et al., 2012; Brondani et al., 2013a), English (Brondani, Luna & Padovani, 2011; Brondani et al., 2013b), Spanish (Brondani et al., 2014), French (Steagall et al., 2017), and Italian (Della Rocca et al., 2018). The instrument is available on the website http://www.animalpain.com.br in Portuguese, Spanish, and English for didactic and scientific training purposes. The UFEPS is divided into subscales such as pain expression, psychomotor changes, physiological variables, and miscellaneous behaviors. The use and application can be time-consuming and complex, and the scale was only validated for cats undergoing ovariohysterectomy (Brondani et al., 2013b). It is not known whether the tool could be applied to other types of surgical or clinical pain. Additionally, it requires blood pressure monitoring which is not always feasible and/or practical in the clinical setting. Recently, a short version of the UFEPS (UFEPS–SF) has been developed in eight different languages to overcome these limitations and facilitate clinical pain assessment in feline practice (Luna et al., 2020). The instrument has easy applicability and has been used in a previous clinical trial (Benito et al., 2019). The UFEPS-SF consists of four items (0–3 points for each item) to evaluate the cats’ posture, activity, attitude, and reaction to touch and palpation of a painful site. The items “appetite” and “blood pressure monitoring” were not included in the short form. The maximum total score is 12, and rescue analgesia is provided at ≥ 4. The instrument and video examples of behaviors of each item may be found at http://www.animalpain.com.br.

According to Merola & Mills (2016), the UFEPS is the “only specific instrument with evidence of validity, reliability and sensitivity at the level of a randomized control trial”. However, if one considers its aforementioned limitations and the lack of another gold-standard instrument for acute pain assessment in cats, the Glasgow CMPS-Feline was used for comparison in the current study because it had been gone through some degree of validation.

The objective of this study was to evaluate the validity (construct and criterion), reliability and sensitivity of the UFEPS and UFEPS-SF, compared with each other and with Glasgow CMPS-Feline for feline acute pain assessment in various painful clinical conditions, and after orthopedic and soft tissue surgery. The hypothesis of the study was that the UFEPS and UFEPS–SF are valid, reliable, and sensitive to the administration of analgesics in cats undergoing different painful clinical and surgical conditions.

Materials & Methods

This was a prospective, clinical, cohort study. It was carried out at the Veterinary Hospital of the School of Veterinary Medicine and Animal Science—São Paulo State University (UNESP)-Botucatu, Brazil, between March and December 2019. The study protocol was approved by the Ethics Committee on the Use of Animals of the same institution under protocol number 0039/2019. Written tutor consent for participation in the study was obtained for each cat.

Pain assessment and rescue analgesia

Three veterinarians evaluated the cats in real-time and in-person: the main observer (MB—completing a MSc. program) and two other graduate students (PhD and MSc) who had previously completed a residency program in Veterinary Anesthesiology (ARO and MTL). Evaluators were trained to use the UFEPS scale by assessing the http://www.animalpain.com.br website to observe the behaviors that corresponded to each item of the scale and by assessing their ability to use the scale. Evaluators did not have access to each one’s scores during evaluation. Initially, they observed the cats from a distance without interaction. Then, the main observer interacted with the animal and performed the physical examination, including pain assessment, and the others only observed. Pain assessment was performed with the unidimensional numerical ratio (NS; 0 “no pain” to 10 “worst possible pain”), simple descriptive (SDS; 1 “no pain” to 4 “worst possible pain”) and visual analog scales (VAS—horizontal line of 10 cm length where “0” corresponds to “no pain” and 10 cm to “worst possible pain”) in this order, followed by the composite scales UFEPS–SF (Table 1), UFEPS (Brondani et al., 2013b) and the Glasgow CMPS-Feline (Reid et al., 2017) in random order. The composite scales are based on several categories, including descriptive levels of behaviors graded from 0 to 3 (UFEPS and UFEPS-SF) or 0 to 4. The six possible orders for assessing the three composite scales were randomized (randomization.org) and excluded after use until all possible orders were used, followed by a new randomization process. The subscale “physiological variables” (i.e., appetite and blood pressure) of the UFEPS was not included in the assessment because repetitive blood pressure monitoring can be stressful in cats, especially in those with painful conditions. In relation to appetite, some cats at some time points could not eat as they were fasted preoperatively. According to the original UFEPS (Brondani et al., 2013b), the subscales may be assessed separately because there is an intervention analgesic point calculated for each subscale. A Portuguese version of the Glasgow CMPS-Feline was translated from English by the authors and used in the study (Reid et al., 2017) and then back-translated by a non-veterinarian independent translator to ensure semantic equivalence. The time to assess each composite scale was recorded in one cat from each group.

Table 1:

Short form of the UNESP-Botucatu multidimensional feline pain assessment scale (UFEPS-SF).

Item	Description	Score
Evaluate the cat’s posture in the cage for 2 min.
1	Natural, relaxed and/or moves normally	0
	Natural but tense, does not move or moves little or is reluctant to move	1
	Hunched position and/or dorso-lateral recumbency	2
	Frequently changes position or restless	3
		Please tick where applicable
2	The cat contracts and extends its pelvic limbs and/or contracts its abdominal muscles (flank)
	The cat’s eyes are partially closed (do not consider this item if present until 1 h after the end of anesthesia)
	The cat licks and/or bites the affected area
	The cat moves its tail strongly
	All above behaviors are absent	0
	Presence of one of the above behaviors	1
	Presence of two of the above behaviors	2
	Presence of three or all of the above behaviors	3
Evaluation of comfort, activity and attitude after the cage is open and how attentive the cat is to the observer and/or surroundings
3	Comfortable and attentive	0
	Quiet and slightly attentive	1
	Quiet and not attentive. The cat may face the back of the cage	2
	Uncomfortable, restless and slightly attentive or not attentive. The cat may face the back of the cage	3
Evaluation of the cat’s reaction when touching, followed by pressuring around the painful site
4	Does not react	0
	Does not react when the painful site is touched, but does react when it is gently pressed	1
	Reacts when the painful site is touched and when pressed	2
	Does not allow touch or palpation	3

DOI: 10.7717/peerj.11225/table-1

Surgery group

The SG cats were admitted 30 to 60 min before the procedure and discharged up to seven hours after surgery. Premedication was performed with methadone (Mytedom®, Cristália, Itapira, São Paulo, Brazil) or methadone and xylazine (Anasedan®, Ceva, Paulínia, São Paulo, Brazil), anesthesia was induced with propofol (Propovan®, Cristália, Itapira, São Paulo, Brazil) alone or combined with ketamine (Dopalen®, Ceva, Paulínia, São Paulo, Brazil) and/or fentanyl (Fentanest®, Cristália, Itapira, São Paulo, Brazil) and maintained with isoflurane alone (Isoforine®, Cristália, Itapira, São Paulo, Brazil) or combined with intravenous (IV) ketamine and/or fentanyl (Table S1). Cats were evaluated immediately before premedication (baseline) and at every hour from 1 to 6 h (or 6.5 h in case cats received dipyrone for rescue analgesia; see below) after extubation (Fig. 2A). The analgesic intervention was performed when the UFEPS-SF score was ≥ 4 out of a total score of 12 points (Benito et al., 2019) using methadone (0.2 mg/kg intramuscularly; IM or IV, if an intravenous catheter was available) and/or dipyrone (Analgex V®, Agener União, São Paulo, Brazil - 12.5 mg/kg; IV) both diluted up to a volume of 1 mL. The choice between the two drugs was based on the observers’ clinical decision and pain intensity. In exceptional cases, when the observers felt that the cats could be in pain, analgesia was provided even if the UFEPS-SF score was <4/12. Pain was assessed 60 and 90 min after administration of methadone and dipyrone respectively, according to the drug pharmacokinetics (Slingsby et al., 2016; Lebkowska-Wieruszewska et al., 2018). If required, the second administration of rescue analgesia consisted of methadone for cats that had received dipyrone and vice-versa. If a third interventional analgesia was required, a combination of methadone (0.1–0.2 mg/kg IM) and ketamine (1 mg/kg IM) was administered. Rescue analgesia was not provided if cats presented signs of dysphoria (restlessness, vocalization, and agitation) within the first hour after extubation. If cats were not painful at 2 or 4 h after surgery, pain assessment at consecutive moments (3 and 5 h) was not performed to minimize the stress of handling.

Results

Forty-three client-owned, mixed-breed cats with clinical conditions or undergoing surgery were enrolled in the study (Fig. 1, Table S1). One feral/aggressive behavior and two cats that left the Veterinary Hospital before the six-hour observational period postoperatively were excluded. Another ten client-owned, mixed-breed, clinically healthy cats without any painful conditions were included for the CG (Fig. 1).

Of all 50 animals recruited, 17 were female (34%) and 33 males (66%), aged 3.8 ± 4.3 years with a bodyweight of 3.8 ± 1.3 kg. Analgesic, anesthetic, surgical procedures, and clinical conditions are described in Table S1.

Duration of pain assessment using the UFEPS-SF, Glasgow CMPS-Feline, and UFEPS (excluding appetite and blood pressure) was 90, 112 and 166 s, respectively. In the OrthG, 11 of 12 cats required the administration of rescue analgesia. In the SoftG, 5 out of 8 cats required rescue analgesia. Seven from 20 cats of the ClinG required procedural sedation and chemical restraint; their pain scores were not included after the administration of sedatives (Fig. 1). Sixteen cats required analgesia in the SG for 31 times (29 times using the criteria of UFEPS–SF scores ≥ 4 and 2 by using clinical judgment).

Discussion

This study showed that the UFEPS and UFEPS–SF are valid, reliable, responsive, sensitive, and specific scoring systems. The study fills a gap by highlighting the reliability and responsiveness of the UFEPS for a pain assessment in cats with different clinical and postoperative pain conditions, including orthopedic surgeries. This overcomes the previous limitation that the scale had been developed and validated only for postoperative pain associated with ovariohysterectomy. In addition, a group of healthy, control cats were included to corroborate the responsiveness of the scale confirmed by finding that pain scores were significantly higher in painful versus pain-free cats. Similar findings were also obtained for the UFEPS-SF. Therefore, this simplified, user-friendly version of the UFEPS can be readily used in feline practice and overcomes another limitation of the UFEPS: the scale is not cumbersome and time-consuming. It can also be used in eight languages (Luna et al., 2020).

The inter-rater reliability of the UFEPS was predominantly good across observers (ICC of 0.79–0.82); however, it was lower when compared to the original study in cats undergoing ovariohysterectomy (ICC of 0.98 for the total score and 0.93 to 0.97 for subscales 1 and 2, interpreted as excellent) (Brondani et al., 2013a; Brondani et al., 2013b). In another study, the inter-rater reliability of the UFEPS for observers with different degrees of experience was moderate (0.7) with great variability in CI (0.2–0.89) (Benito et al., 2017), suggesting that training may affect the reliability of these pain scoring systems. On the other hand, the inter-rater reliability of the Glasgow CMPS-Feline and UFEPS-SF was moderate to good (0.65–0.82 and 0.63–0.82, respectively). We hypothesize that these scales have less detailed descriptors and worse reliability when compared with the UFEPS. However, this should not have a major clinical impact; these instruments can be used in different pain conditions if they are used by individuals with experience in pain assessment.

Inter-rater reliability was moderate for the unidimensional scales (NS, SDS, and VAS). Similar ICC results were found when a dynamic and interactive VAS was used by observers with different experiences (Benito et al., 2017). However, their CI values were lower (0.19 to 0.8) compared to the current study (0.57 to 0.78). Unidimensional pain scales may not capture all the complexity of pain (Robertson, 2018). Nevertheless, they might be acceptable for clinical pain assessment when used by experienced observers.

The correlation of a new scale with instruments with reported validation (i.e., gold-standard) is required in the study of concurrent criterion validity (Streiner & Norman, 2008). The correlation of the UFEPS–SF with the UFEPS and Glasgow CMPS-Feline was very strong and strong, respectively, and confirms the criterion validity for this scale. In previous studies, criterion validity was very strong between the Feline Grimace Scale and the Glasgow CMPS-Feline (0.86) without the facial component (Evangelista et al., 2019) and strong between the UFEPS and the Glasgow CMPS–Feline (0.6–0.8) (Steagall et al., 2018). The criterion validity between the UFEPS and UFEPS-SF with the Feline Grimace Scale should be a subject of a future study.

The ability to detect a significant change in pain scores, whether by decreasing the score after rescue analgesia or increasing the score after a painful procedure, is part of construct validity or responsiveness of the instrument (Von Baeyer & Spagrud, 2007; Chien et al., 2013). It determines whether the scale can detect differences between known groups (e.g., painful versus non-painful individuals) (McDowell, 2009; Brondani et al., 2013b). Such comparisons have already been used in the validation of the Feline Grimace Scale and the Glasgow CMPS-Feline (Reid et al., 2017; Evangelista et al., 2019). The construct validity of the UFEPS was previously determined by comparing baseline scores with the highest pain postoperative scores (Brondani et al., 2013b). The absence of a negative control group in Brondani et al. (2013b) study was a limitation described in a systematic review (Merola & Mills, 2016). All scales used in the current study, including the UFEPS, distinguished animals with clinical and surgical pain from those without pain, demonstrating the responsiveness of the UFEPS and UFEPS-SF for both the administration of rescue analgesia and in comparison, with controls. The limitation of the construct validity in the current study was that the observers were biased to the painful status of these cats. Indeed, they knew if cats had had surgery or required analgesic administration for pain relief. The same occurred for the control cats; the observers knew that these cats were most likely not suffering from any painful condition. Therefore, the scores given for each scale could be influenced by the assessments carried out previously. These limitations may have inflated the scores given before rescue analgesia and deflated the scores given after administering analgesia independently of the scale used. Both clinical real-time and image assessment of the Feline Grimace Scale (Evangelista et al., 2020) and video scoring with the UFEPS (Brondani et al., 2012; Brondani et al., 2013b) was reported in previous studies. Otherwise, construct validity of the UFEPS–SF should be further corroborated in a future study via video assessment with observers who are blinded to the analgesic administration and painful status.

Regarding sensitivity, the UFEPS–SF and Glasgow CMPS-Feline scales detected the most truly painful cases. Overall, the sensitivity of the unidimensional and multidimensional scales was moderate, showing that close to 80% of the cats suffering pain would be correctly diagnosed (true pain). Specificity showed that the three multidimensional scales detected most of the true negatives; the UFEPS-SF had the best specificity to identify pain-free cats. Although the Glasgow CMPS-Feline had similar results to the other multidimensional scales, it presented lower sensitivity and specificity than the other scales. One possible reason was the translation into Portuguese; however, back translation of the instrument ensured semantic equivalence. Translation and back-translation of a scale are required to validate a scale to be used in a different language. It is important for the semantics and terminology of the new instrument (Streiner & Norman, 2008; Sousa & Rojjanasrirat, 2011). Likewise, for responsiveness, the results for specificity were also biased in this study since observers were most likely aware of the cats’ painful status. Therefore, it is not surprising that the unidimensional scales had an excellent specificity (i.e., detection of true negatives; non-painful client-owned cats from the Veterinary Hospital personnel).

A limitation was that the results for the unidimensional scales might have been influenced by not randomizing them before the assessment and not performing their evaluations individually. Another limitation was the order of pain assessment; the scores of the unidimensional scales might have influenced the scores given in the subsequent composite scales. The authors decided to prioritize the assessment of the composite scales because this was the primary objective of the study. Pain assessment using the composite scales could have inflated the scores of the unidimensional scales had they been used first because they indicate which pain behaviors should be assessed (Roughan & Flecknell, 2006). Considering that unidimensional scales are subjective because they do not include pain behaviors in their assessments, the authors considered that unidimensional instruments would have less influence in subsequent composite scale assessments than the other way round. This approach was similar to other previously published papers in cats and other species (Brondani et al., 2013b; De Oliveira et al., 2014; Silva et al., 2020).

A possible confounding factor that may influence postoperative pain assessment is the use of sedatives and anesthetics. Indeed, pain could not be assessed in some cats one hour after the end of surgery due to the presence of residual anesthesia. Pain could be overestimated with false-positive results, as described anecdotally in dogs (Mathews et al., 2014). Under these circumstances, cats would receive unnecessary analgesia. The same applies to postoperative dysphoria and excitement. Most of the cats received ketamine for induction of anesthesia and some for intraoperative pain management. Ketamine increases psychomotor scores using the UFEPS, falsely increasing pain scores (Buisman et al., 2016). Another limitation of the study is that only animals presenting feral or aggressive behavior were excluded. Shy individuals were included. Shy and aggressive cats may have higher scores on the psychomotor subscale of UFEPS and Glasgow CMPS-Feline scales (without the facial component) (Buisman et al., 2017) due to their unique demeanor.

It is not known if pain scores were affected by the presence of observers and potential cat owners in this study. In a previous study (Evangelista et al., 2020) for the Feline Grimace Scale, there was no significant difference between real-time and video assessments. Real-time scores were slightly overestimated when compared with video scores, which would probably not affect the clinical assessment. These are some limitations that demonstrate the challenges of clinical pain assessment and the development and validation of pain scoring instruments in veterinary medicine.

In summary, possibly because unidimensional scales have no descriptors, they have less inter-rater reliability than the composite ones. However, concurrent criterium validity, responsiveness, sensitivity, and specificity were comparable for both unidimensional and composite scales. The UFEPS-SF had the lowest possibility to have pain-free cats diagnosed as painful and provided the quickest assessment time when compared to others. It is important to highlight that training was performed beforehand and that the robustness of theses scales should be assessed using untrained observers.

Conclusions

Both UFEPS and UFEPS–SF showed appropriate concurrent validity, reliability, responsiveness, sensitivity, and specificity for feline acute pain assessment in cats with various clinical conditions and those undergoing orthopedic and soft tissue surgery. The results of this study for the UFEPS-SF should be corroborated in a future study by using a masked and randomized design.

Supplemental Information